Breaking: remove methods that are hardly uesd and cause many invalidations (#1113)

* do 0.6, 0.7.2 broke for us (#1099)

* Fix tests on julia 1.12 (#1110)

* use isequal instead of === to compare NaN

* drop all and broadcast

* specify DimensionalData.Dimensions to make reference unique in docs

* drop convert method for name to abstractstring

* remove `merge` method for dimstack with iterators of pairs

* add to changelog

---------

Co-authored-by: Lazaro Alonso <[email protected]>

Author: tiemvanderdeure

CHANGELOG.md

@@ -40,7 +40,7 @@ Changelog.generate(
 
 - `similar` for `AbstractDimStack` has modified behavior more consistent with
   `similar` for `AbstractArray` ([#903]).
-- Internal const `CategoricalEltypes` now includes `DataType` - sot that a
+- Internal const `CategoricalEltypes` now includes `DataType` - so that a
   lookup of types is considered `Categorical` by default ([#876]).
 - `skipmissing` on an `AbstractDimStack` now skips any `missing` values in any fiels, rather than `missing`, which can't
   actually occur ([#1041]).
@@ -51,12 +51,15 @@ Changelog.generate(
     DimArray{T}(x::UndefInitializer, dims::MaybeDimTuple; kw...) where T = DimArray(Array{T}(undef, map(length, dims)), dims; kw...)
     MyDimArray(st::AbstractDimStack; kw...) = DD.dimarray_from_dimstack(MyDimArray, st; kw...) 
 - the deprecated `index` function is now removed completely
-- `metadta(obj, dims)` no longer works as its ambiguous and not consistently
+- `metadata(obj, dims)` no longer works as its ambiguous and not consistently
   implemented - `metadata(obj)` does not return a Tuple for each dimension 
   like other similar methods, but the metadata of `obj`. To get dimension 
   metadata going forward, explicitly use `metadata(dims(obj, X))`
-- `val(obj)` no longer returns `map(val, dims(obj))` because thats just weird,
+- `val(obj)` no longer returns `map(val, dims(obj))` because that's just weird,
   val seems like it would return `parent` from that call. `lookup` does this anyway.
+- Removed two methods that don't have much use and cause lots of invalidations:
+  - `Base.convert(Type{<:AbstractString}, name::DD.AbstractName)` is removed, use `string(name)` instead.
+  - `Base.merge(stack::AbstractDimStack, pairs)` is removed, use `merge(stack, NamedTuple(pairs))` instead.
 
 
 <!-- Links generated by Changelog.jl -->

docs/package.json

@@ -6,7 +6,7 @@
     "markdown-it": "^14.1.0",
     "markdown-it-mathjax3": "^4.3.2",
     "vitepress": "^1.6.3",
-    "vitepress-plugin-tabs": "^0.7.1"
+    "vitepress-plugin-tabs": "^0.6.0"
   },
   "scripts": {
     "docs:dev": "vitepress dev build/.documenter",

docs/src/dimensions.md

@@ -36,7 +36,7 @@ DimensionalData.jl uses `Dimensions` everywhere:
 This symmetry means we can ignore how data is organized, 
 and label and access it by name, letting DD work out the details for us.
 
-Dimensions are defined in the [`Dimensions`](@ref) submodule, and some 
+Dimensions are defined in the [`DimensionalData.Dimensions`](@ref) submodule, and some 
 Dimension-specific methods can be brought into scope with:
 
 ```julia

src/name.jl

@@ -5,8 +5,6 @@ Abstract supertype for name wrappers.
 """
 abstract type AbstractName end
 
-Base.convert(T::Type{<:AbstractString}, name::AbstractName) = convert(T, string(name))
-
 """
     NoName <: AbstractName
 

src/stack/stack.jl

@@ -241,9 +241,6 @@ function Base.merge(
 )
     rebuild_from_arrays(x1, merge(map(layers, (x1, x2, xs...))...); kw...)
 end
-function Base.merge(s::AbstractDimStack, pairs; kw...)
-    rebuild_from_arrays(s, merge(layers(s), pairs); refdims=())
-end
 function Base.merge(
     x1::NamedTuple, x2::AbstractDimStack, xs::Union{AbstractDimStack,NamedTuple}...;
 )

test/groupby.jl

@@ -100,9 +100,9 @@ end
             mean(A[X=x, Ti=d])
         end
     end
-    @test all(collect(mean.(gb)) .=== manualmeans)
-    @test all(mean.(gb) .=== manualmeans)
-    @test all(combine(mean, gb) .=== manualmeans)
+    @test isequal(collect(mean.(gb)), manualmeans)
+    @test isequal(mean.(gb), manualmeans)
+    @test isequal(combine(mean, gb), manualmeans)
 end
 
 @testset "broadcast_dims runs after groupby" begin

test/name.jl

@@ -7,13 +7,11 @@ using DimensionalData: Name, NoName
     @test Symbol(Name(:x)) === :x
     @test string(Name(:x)) === "x"
     @test Name(Name(:x)) === Name(:x)
-    @test convert(String, Name(:x)) === "x"
 end
 
 @testset "NoName" begin
     @test Symbol(NoName()) === Symbol("")
     @test string(NoName()) === ""
     @test Name(NoName()) === NoName()
-    @test convert(String, NoName()) === ""
 end
 

test/stack.jl

@@ -147,7 +147,6 @@ end
     @test merge(mixed) === mixed
     @test keys(merge(mixed, s)) == (:one, :two, :extradim, :three)
     @test keys(merge(s, mixed)) == (:one, :two, :three, :extradim)
-    @test keys(merge(s, (:new=>da4,))) == (:one, :two, :three, :new)
 end
 
 @testset "setindex" begin